Utility Maximization in Satellite Networks Using Onboard Caching

Hao Wu,Jian Yan,Linling Kuang
DOI: https://doi.org/10.1109/ICTC57116.2023.10154792
2023-01-01
Abstract:The predictable mobility of satellites enables a store-carry-forward paradigm for flows tolerant of long delivery time. Compared to end-to-end direct transmission, this paradigm has less inter-satellite link usage. Using onboard caches to store and carry flows, we propose a time-expanded graph (TEG)-based traffic engineering (TE) scheme. We use a conventional resource allocation algorithm for single-epoch network utility maximization. Then, we propose a load-and-delay-aware path selection (LDAPS) method to control multi-epoch resource contention and improve long-term network utility. Simulation of an actual satellite system reveals that our scheme achieves higher long-term network utility than existing non-caching TE methods. Under the same cache size, LDAPS has better performance than the shortest path strategy.
What problem does this paper attempt to address?